The racial homeownership gap is real. It impacts economic opportunities, embeds city and neighborhood segregation, and is a big contributor to wealth inequality. In Oregon, Black, Indigenous, and People of Color have an overall homeownership rate that is 15 percentage points lower than their white, non-Hispanic neighbors. The little bit of silver lining is recent homeownership gains are more broad based and housing wealth is becoming a bit more equitable, but there is still a long way to go.
This edition of the Graph Table of the Week is not exactly about this very real issue, but about one way we talk about it. Earlier this year there was a report on the racial homeownership gap across states. It showed that Oregon’s gap is smaller than in a lot of other states. On the surface that is good news. But in digging into the data it reveals something more important to note. It reveals a Simpson’s Paradox which is when topline level data masks what is actually happening.
First it must be noted we are still dealing with lagged Census (ACS) data. The most recently available data is for 2021. We get the 2022 ACS data this fall (published tables in September).
Second, on the surface the racial homeownership gap in Oregon is smaller than it is nationwide. Oregon’s 15 percent gap is less than the 22 percent gap across the country. This relatively smaller gap is for two reasons. First, the white, non-Hispanic homeownership rate in Oregon is much lower than it is nationally, primarily due to our bad housing affordability that puts ownership out of reach for some households. Second, the BIPOC homeownership rate is slightly higher than it is nationally, thus showing a narrower spread overall.
But the reason why Oregon’s BIPOC homeownership rate is slightly higher is where it gets more informative. If we look at homeownership rates by each major race or ethnicity it shows that Oregon is below the national figures pretty much across the board. So if we are lower for each race or ethnicity individually, how can we be higher for our BIPOC homeownership rate overall? That’s the Simpson’s Paradox, and you can see the answer in the set of numbers over on the right.
Among BIPOC Oregonians, our demographic composition is significantly different than the nation. The biggest factor here is Oregon’s small Black population compared to the nation, and the fact that the Black homeownership rate is the lowest. These two factors combined shift the overall composition and calculation when looking at BIPOC overall. In a mathematical sense, Oregon’s BIPOC homeownership calculation is weighted more towards our relatively larger Asian and Hispanic population which have somewhat higher ownership rates, and weighted less toward our relatively smaller Black population and lower ownership rates.
Now, you can re-run the calculations and control for the demographic differences. If Oregon’s racial and ethnicity makeup were to match the nation, our BIPOC homeownership rate would be 48.6 percent, which would be 2.5 percent lower than the national BIPOC ownership rate. That would still put Oregon’s racial homeownership gap somewhat smaller than the US (19 percent versus 22 percent), but on a proportionate basis it would be pretty similar given Oregon’s overall lower homeownership rates.
First, we know there was a household formation boom earlier in the pandemic. The increases were especially strong in the ownership market. On one hand that means Oregonians, like Americans in general, loaded up on mortgage debt. From 2019 to 2022, mortgage debt for Oregonians increased 27% and now stands are nearly $184 billion accordingly to the recently released data from the NY Fed. However, on the other hand, it’s never just about the level of debt itself, but also the value of the asset and the ability to service the debt. For instance, personal income in Oregon increased 18% from 2019 to 2022, and that excludes the recovery rebates and other pandemic aid programs that boosted household finances even more. As such, mortgages as a share of income only ticked up modestly.
Quick note: When it comes to household debt, the vast majority is mortgages. Homes are the biggest investment most households ever make. Given the debt is also related to prices and affordability, Oregonians mortgage debt as a share of income ranks 8th highest nationwide. When it comes to other household debt like autos, credit cards, and student loans as a share of income, Oregon ranks 28th highest.
Second, it’s not just about the income needed to service the debt, but also the interest rate paid. We had record low mortgage rates earlier in the pandemic. The actual mortgage payments U.S. households are making when measured as a share of disposable personal income are lower than they were pre-pandemic. This is a combination of new borrowers buying, and existing homeowners refinancing. While the data at the link above is national, we do know Oregonians followed similar trends. The latest data from the FHFA shows that 68% of Oregonians with a mortgage have an interest rate below 4% on their home.
Looking forward these dynamics have changed. Obviously mortgage rates have spiked and are now in the high 6 percent range lately. When it comes to the mortgage rate lock idea that households do not want to give up their low rate and buy at today’s higher rates, it really comes down to overall affordability and not purely the rate itself. And overall homeownership affordability at the moment is terrible. But, we continue to see some sales. According to RMLS, home sales in the Portland market are down 30-40%. They are not down 100%. That means there are buyers at today’s prices and rates, just fewer of them. The market is adjusting. We will get the FHFA data for 2023q1 here in a few weeks and will be interesting to see how much of an increase there is among those with a greater than 6 percent rate.
Even so, we know resale inventory remains at low levels. Fewer households than usual are looking to move, likely due to the affordability shock, sometimes referred to as the seller’s strike. What this means is that new construction becomes even more important. New construction is an increasing share of the for sale market when resale inventory is low.
Third, and finally, when it comes to debt we have to take into account the value of the asset as well. With home prices, and the homeownership rate increasing in recent years, our office’s estimate of owner-occupied home equity reached a new high in 2022 when measured as a share of the economy. Now, with prices down a few percentage points (the numerator), and continued underlying economic growth (the denominator), we are seeing a slight step back in 2023 and in 2024, before continued gains in the years ahead based on our latest forecasts for households, home prices, and the economy.
While all of this is undoubtedly good news for current and future homeowners, we know it’s not from a housing market perspective. Much of this wealth is inequitably distributed (although a bit less today than in years past) and driven by scarcity and our unwillingness to build enough housing. Now, homeownership is good in the sense of the forced savings, and paying down debt perspective, plus locking in a fixed monthly payment and the like. But, housing cannot be both a good investment and affordable.
Another quick note. One of the reasons why equity is rising is due to households paying off their mortgages. I have seen some national figures indicating 40% of US homeowners do not have a mortgage. At least in 2021, we know 35% of Oregon homeowners do not have a primary mortgage (We get 2022 ACS data in September). Given demographics, the absolute number of owners without a mortgage should continue to increase. Such households, if they choose to move, are more likely to be cash buyers when the sell and buy again, or can tap the equity with a HELOC and the like.
If we step back and look at the economic cycle to date, it is clearly different than others in recent decades. Typically, Oregon is more volatile than the nation and the typical state. While recessions and expansions locally are perfectly timed with the nation — we tend to neither lead nor lag — the downturns are deeper, and the expansions faster. So far that has not been the case this cycle and Oregon currently ranks in the middle of the pack across all states. Part of this is due to the synchronized nature of the pandemic, the shutdowns, the federal aid and the like. Part of this is also probably the point in the cycle we are today, being just 3 years into it, there hasn’t been a lot of time for more variation to develop around the country. But part also seems to be tied to a relative shift in the composition of growth in recent years.
First, we included this chart in our presentation to the Legislature last week, and here on the blog, but I wanted to expand on it a bit further today.
Local employment trends are a bit below average. The change in employment since the start of the pandemic ranks 30th fastest out of all states, and trails the nation by one percentage point. The most recent round of annual benchmark revisions from the U.S. Bureau of Labor Statistics widened this gap as Oregon saw a slight downward revision to the published data in the past year. Examining trends across sectors, Oregon’s relatively slower employment growth compared to the nation is primarily due to slower recoveries in health care and leisure and hospitality, while stronger gains in construction help offset some of the weakness.
More encouraging are local income trends that are a bit above average. Total personal income in Oregon has increased 19.8 percent from the end of 2019 through the end of 2022. This large increase translates to the strong increases in consumer spending and income taxes paid in recent years. Of course the high inflation during the pandemic means the cost of living has also increased during the same time period. After adjusting for inflation, total Oregon personal income is 3.5 percent higher than at the end of 2019 while the U.S. is up 2.6 percent. Oregon’s income growth ranks 21st fastest across all states. This relative strength is primarily in the non-wage forms of income, which also have an outsized impact on tax revenues.
While Oregon is right in the middle of the pack when it comes to employment and income, that same cannot be said for local GDP where the state has experienced the 11th strongest growth across the country. Local GDP data is prone to revisions and assumptions about productivity and value-added. As such it may be best to wait for a few more quarters of data, and the upcoming comprehensive revisions this fall to know exactly where the state stands, even as it is encouraging to see the strong growth.
What is most interesting overall is the composition of this economic growth is a bit different than what Oregon has experienced in recent decades. Today employment and population gains are lagging the nation. Instead, it is income and production leading the recovery. In much of the 1990s through early 2010s it was Oregon’s relative income growth that was most concerning as it lagged, while jobs were plentiful and grew at a fast pace. This relative pattern of growth started to shift late last decade and appears to have continued through the pandemic and recovery. To a certain extent it does not matter the exact composition of economic growth, but it is determined by both the amount of capital and labor in the region. We talked more in-depth about the labor force outlook the other week, and there is plenty more on that in our latest forecast doc.
Capital and Productivity
Increased productivity raises the overall speed limit of the economy. Producing more per hour work increases business revenue and worker wages. However, it typically takes investment in the various forms of capital on the part of both businesses and workers to raise productivity. Capital can take different forms includes financial, physical, natural, human, and social. No firm or region excels at or has access to each type of capital. However, they can rely more upon the other types of capital that it does have for future growth.
Overall Oregon has great productivity. Part of this is the state’s historical strength in high-tech manufacturing which is a highly productive industry. But improvements in productivity in recent decades are broader than that. In the real GDP estimates currently published, it is not Oregon’s manufacturing output per worker driving the above-average gains, but rather strength across a variety of industries.
The U.S. Bureau of Labor Statistics also publishes state level estimates of labor productivity. Last cycle, from the height of the housing boom in 2007 through the pre-COVID peak in 2019, Oregon’s labor productivity ranked 5th fastest among all states. From 2019 to 2022, Oregon’s labor productivity ranked 3rd fastest among all states.
Let’s focus on the labor productivity numbers for a minute because BLS just released the 2022 data this morning, along with some revisions*. There are a handful of important things to keep in mind. First, productivity surged early in the pandemic. This was not really because of new technologies or increased efficiencies, but rather because there were a lot of people out of work, yet consumer demand remained strong, and so businesses had to make do with the workforce they had. Output per worker soared well above the pre-pandemic trend, but it was sort of a forced productivity increase, if you will.
Second, in 2022, Oregon’s labor productivity declined 1.9 percent. This decline was not unexpected. The national figures had already shown some declines, and now the state figures do as well. The big reason here is that employment and hours worked grew at a faster pace than real value-added output did last year. Essentially payrolls ramped back up and even as the economy grew and output grew, some of those initial/forced productivity gains proved temporary. Plus, it takes time for new workers to gain experience and become more productive in their jobs. Even so, Oregon’s labor productivity in 2022 is 9 percent higher than in 2019 (3rd largest increase nationally), and 3.5 percent above the pre-pandemic trend.
Looking forward, it has to be noted that economists struggle with forecasting productivity gains. It’s not just about the number of firms and workers, and the level of business investment, or federal funding for R&D, but whether or not all of those things mix together into something that makes the economy more efficient. That’s hard to know, and in recent decades productivity has been slow. Economists tend to think about start-ups and new technologies being key catalysts for productivity gains. Which is why, as we have discussed previously, productivity may be stronger in the years ahead due to the increase in new business formation. New firms typically bring new ideas and products, and improve efficiencies compared to existing firms, which raises economywide productivity. But the gains also come from existing firms as they do their own R&D, and try to improve business processes and the like.
Finally, when it comes to Oregon’s relative growth patterns, it can be hard to know how much of the current data is more a moment in time versus a true, fundamental shift. I think that largely comes down to migration and population growth. If migration rebounds as expected, then Oregon will likely continue to rank well in terms of population and employment growth, to go alongside of the need for capital investment to grow the economy. However, if migration does not rebound as expected, then Oregon will need to rely even more on being a place business want to invest, which in turn will drive productivity gains, business revenue, and worker wages.
* In our forecast document we noted Oregon ranked 4th fastest from 2007 to 2019, and in the newly revised data Oregon ranks 5th. We also wrote that from 2019 to 2021, Oregon ranked 6th fastest, but in the newly revised data and adding in 2022, Oregon ranks 3rd fastest from 2019 to 2022.
This morning the Oregon Office of Economic Analysis released the latest quarterly economic and revenue forecast. For the full document, slides and forecast data please see our main website. Below is the forecast’s Executive Summary and a copy of our presentation slides.
Inflationary economic booms have not traditionally ended well, meaning not without a recession. As such it is easy to be pessimistic about the outlook for the economy. Economic developments like last year’s goods recession, and the banking turmoil earlier this year add more fear to the outlook. However, a near-term recession is far from a slam dunk. The reasons include some nascent signs that inflation is cooling and the Federal Reserve is looking to pause its interest rate increases which limits the potential for overtightening. Furthermore, the economy is showing some signs of renewed strength as housing and manufacturing stabilize, and income growth is again outpacing inflation. All of these indicate a sudden stop in the economy in the short-term is unlikely. Part of forecasting is not just identifying the dynamics, but also the timing.
Our office’s baseline forecast calls for the economic soft landing and continued expansion. To be sure, recession risks remain very real, but given the current economic dynamics, those recession risks are likely a 2024 or beyond story. Even so, Oregon’s economy will slow noticeably in the upcoming 2023-25 biennium, however for good reasons. The recovery from the pandemic has been faster, and more inclusive than any in recent memory. With the economy operating at or near full employment, underlying gains in the labor market will be closely tied to demographics and population growth. To maintain even stronger economic growth in the years ahead Oregon will need to see faster population gains, and/or rely on business investment and capital to increase productivity. This cycle has been different in the sense Oregon ranks in the middle of the pack economically with income and productivity outpacing the typical state while jobs and population lag, the opposite pattern of decades past.
Available resources are expected to be up sharply relative to what was assumed in the March 2023 forecast, both in the near term and over the extended horizon. The upward revision in the outlook is based both on a stronger than expected tax filing season, as well as methodological changes made in light of fundamental shifts seen in recent years.
The tax filing season once again outstripped expectations, albeit modest ones. Revenue gains have cooled some, but it is clear that Oregon’s tax sources have become more effective than they were pre-pandemic.
One major factor has been the current inflationary environment. The vast majority of Oregon’s taxes are not adjusted to inflation and rise along with prices. With demand outstripping supply, businesses and consumers are paying premiums for their needs. This has translated into a wide range of taxable business and labor income, which has moved many filers into higher tax brackets. The new Corporate Activity Tax, Vehicle Privilege Tax, alcohol, and tobacco taxes have risen with inflation as well.
Inflationary dynamics have not been captured well by Oregon’s revenue models, given that this sort of environment has not existed since years before computerized models have. Oregon’s revenue models have also been refined to better account for fixed tax brackets and federal tax reform.
Qualitatively, there is not much difference between the updated revenue outlook, and what was predicted in March. After unsustainably high revenue collections over the past two years, tax revenues are expected to come back to earth over the next biennium, before returning to healthy growth thereafter.
Quantitatively, small differences in trajectory matter a lot, and compound over time. Taken together, the outlook for personal and corporate income taxes has risen by $1.5 to $2 billion over the forecast horizon due to the updated model methodology. The 2021-23 personal kicker is now estimated to be $5.5 billion, and the corporate kicker is now estimated to be $1.8 billion.
See our full website for all the forecast details. Our presentation slides for the forecast release to the Legislature are below.
Last week we talked a bit about the labor force outlook and how Oregon appears to be at or near its potential given the strong cyclical economy and aging demographics. Besides labor, the other major driver of economic growth is productivity, which requires business investment and gains in the various types of capital: financial, human, natural, physical, and social. Historically a key source of these gains comes from new businesses. Generally new firms bring new ideas, and/or improved processes that are more efficient to the marketplace. Older or more rigid businesses get left behind as part of the creative destruction process. Start-up activity had clearly been trending down for decades leading up to the pandemic. Since then, new business formation has increased noticeably and appears to be sticking, at least so far as seen in the latest edition of the Graph of the Week.
Running a business is hard. Many firms do not make it beyond their first year or two. Only a couple eventually grow to become a huge, successful firm. As such, one way our office thinks about this is not at the individual firm level, but for Oregon’s economy overall. Even if the probability that any given business will be successful remains the same, if you have more ping pong balls in the hopper, your overall probability that one of them will be successful is higher. And while simply having more businesses does not necessarily lead to increased business investment and productivity gains, it sure is an encouraging signal about those possibilities in the years ahead.
In terms of the outlook for start-up activity there are some crosscurrents. Personal savings and home equity are the most common funding source for new businesses. While those are higher today, and can continue to support and drive this ongoing strength in start-up activity, the impacts of high inflation, rising interest rates, and recession risks likely weighs on new activity a bit. Additionally, we know venture capital and banking lending is tighter today than earlier in the pandemic, which will likely slow new business formation as well.
Now, on the upside two items stand out. One is the potential for a general shift in the economy, especially related to working-from-home type jobs and businesses, which allows more people to test out their ideas if they so choose. Two is demographics. While young college dropouts may be the stereotype of who are entrepreneurs, most are actually middle-aged. Research from the Census Bureau finds the average age of the entrepreneur at founding is 42 and the average age of the most successful of the new businesses is 45. Obviously, there is a range of ages and outcomes, but on a demographically-adjusted basis entrepreneurship rates peak in the late 30s and early 40s. The fact that the large Millennial generation is now in their 30s and increasingly in their 40s, that is a big demographic tailwind to new business formation. A key here is that entrepreneurs tend to get a bit of work experience and in doing so they identify a need or a new way to do things that is more efficient, and then go out and try to improve that thing they have identified, and hopefully make money along the way.
Finally, as noted at the top, if we think about the capital and labor as the nuts and bolts of growth, there are a few implications given we have a pretty good idea where labor stands today. It means attracting and retaining talent is even more important (and getting the expected rebound in migration is huge for the long-run outlook). It also means capital takes on extra importance, and in terms of the range of possible outcomes, it is more of a wildcard at the moment.
As Oregon’s economy has recovered from the pandemic, job growth is starting to slow. A key concept in economics is that of full employment, or the natural rate of unemployment, sometimes referred to as NAIRU or u* (pronounced u star). Basically it’s the highest employment level an economy can have without generating higher inflation and when nearly everyone who wants a job, has a job. I know what you’re thinking. With today’s high inflation that must mean we are beyond full employment. From a demographic and employment view I don’t think that’s correct, but I do think we have seen more wage-price persistence that many economists expected. A common view today, held by our office as well, is that as the labor market cools just a bit, it will lessen the pressure on the market and wage growth will slow back to something more consistent with the Fed’s inflation target. Tomorrow we get the 2023q1 employment cost index reading for the nation which is the best measure of wage growth. It started to slow at the end of last year, and we will see if that continued or not.
With that preamble out of the way, a key number I’ve been watching closely this year is Oregon’s labor force participation rate. Last month Oregon’s LFPR stood at 62.7%. At first blush that can seem low. But it’s important to remember that the standard LFPR is calculated based on the entire population 16 years and older. That means our shifting demographics, where the large Baby Boomer generation is retiring, mathematically pulls down the labor force participation rate even if younger Oregonians seek work at historically normal rates. As such, March’s 62.7% labor force participation rate in Oregon is almost exactly what one would expect in a full employment economy given today’s demographics.
Now, that demographically-adjusted full employment participation estimate is anchored to what the labor market looked like in 2000, the last time we know the U.S. economy was truly at full employment, if not a bit beyond it. This is common technique to use given the lackluster and largely incomplete economic recoveries this millennium. As seen in the chart, the common state of the US economy in recent decades has been underperforming. Of course this is all old news. But it raises an important question. Should we still be thinking about 2000 as a proxy for what we think the labor market should look like?
Given we know today’s labor market is very strong as well, does it make sense to update our views and think about now as the new normal, or the new full employment? Here’s what I mean. Let’s take a look at the labor force participation rate by age today and compare it to 2000. I think this chart is informative.
First for most prime working-age cohorts today’s LFPR is the same as back in 2000, the vast majority of 30-54 years today work just as much as their counterparts did during the tech boom. What’s different however is lower participation rates today among teenagers and 20-somethings, but a corresponding increase among 60- and 70-somethings.
The question is how should we think about this? We have noted before the increase in employment among 60- and 70-somethings. I think that is generally here to stay. It’s in part due to more office-based work and less manual labor allowing workers to physically work longer if they so choose, and in part financial where some of us have to work longer because we do not have enough retirement savings. On the other end, much of the decline in participation among younger Oregonians (and Americans) is due to increased schooling. And increased schooling may depress participation and the labor force today but *should* increase wages and hopefully productivity in the future. But the increased schooling was a 2000s and 2010s story and is not a pandemic story when we know higher ed enrollments are down. Nationally, participation rates among teenagers is rising slightly, but has a long way to go to get back to 1990s readings. And 20-24 year old participation is better described as flat, although ticking up in recent months.
These differences can matter when thinking about the outlook and what Oregon’s potential labor force, and potential economic growth could look like. Now, if we take the 2000 age profile of the labor force, and the 2022 age profile of the labor force and compare them, they look pretty similar in aggregate. See the two solid gray lines in the chart below. The 2022 profile is a bit stronger given current demographics where the higher participation rates among 60-somethings matters more given the Baby Boomers and the lower participation rates among younger Oregonians matters less given the smaller Gen Z population.
But it’s really that top, dashed line that I am curious about. All that does is take the maximum LFPR by age from either 2000 or 2022 and maps that to our demographic forecast in the decade ahead. What it shows is if Oregon were to maintain the higher participation rates among older workers and a return to higher rates from younger workers at the same time, Oregon’s potential labor force would be at least 100,000 higher. That’s a huge number, especially when we have more job openings today than unemployed Oregonians.
Now, I don’t know that it is a reasonable assumption to bank on suddenly reversing 20 year trends (lower participation among younger workers). But thinking through exercises like this is important when assessing where we think the economy is going or could potentially go in the years ahead. Remember, economic growth is about how many workers you have and how productive each worker is. What this shows is we need to add younger workers to the list of potential pockets of labor that businesses need to tap into more if they are looking to hire and expand, or even just replace retirements. This would be in addition to the Latent Labor Force which is about the possibility of reducing historical disparities when it comes to differences based on sex, educational attainment, and race and ethnicity.
Finally, this baseline work is also important when thinking about alternative scenarios of the future could look like. For instance, if population growth and migration does not rebound as expected, it means local firms will need to tap into the existing pool of labor to a greater degree to hire. And if we look across the nation, some of the places with higher employment and participation rates on a demographically adjusted basis are places like the Midwest which hasn’t seen the population and migration trends that the West and South have in recent generations. Our office will have more to say on these potential possibilities at a future date.
Today is April 20th and there has likely never been a better time to be a cannabis consumer in Oregon. Prices are at record lows, having fallen significantly in the past couple of years. Given the large number of retailers in the state, product is widely available at these low prices. Specifically, wholesale prices for usable marijuana are down 54%, while retail prices are down 30% since March 2021. Extract prices are down 15% and are also at record lows.
While the oversupply of production and saturated retailer market that continue to drive prices lower is not new news, these ongoing issues have really come to a head in recent quarters.
First, today’s lower prices do not appear to have resulted in an increase in quantities sold. Now, OLCC estimates that the total amount of THC sold in 2022 increased compared to 2021 but at the product level the number of pounds of usable marijuana, or the number of edibles and the like is steadier. Consumers in the past year appear to have stable consumption patterns and have pocketed the savings or had to spend it on other items in their budget due to high inflation.
Second, if the volume of sales is relatively steady but prices are declining that means business revenues, and tax collections are declining as well. It’s a complicated picture of businesses struggling with market conditions and being unable to pay all their bills. As firms struggle with profitability, we have seen a rise in tax delinquencies. The cascading impact is for those lower on the priority list, be they growers on consignment or the taxing authorities to see the biggest impact. Now, our friends over at the Oregon Department of Revenue are of course working with firms who are behind on their taxes, as they always do with taxpayers and through enforcement activity some revenues are likely to be regained.
The bottom line for the outlook is our forecast has been lowered both today and over the entire 10 year forecast horizon due to these dynamics. Specifically, resources in the current 2021-23 biennium have been lowered by $25 million in the past two quarterly forecasts. More than half of this reduction is due to lower-than-expected tax payments relative to actual sales. Now, sales themselves have been a few percent lower than previously forecasted, but even after taking this reduction in the outlook, recent tax collections are even weaker than that. Looking forward, sales and tax revenues will grow, however all future biennia’s forecasts are revised downward by 5-10 percent relative to previous expectations.
I want to circle back on the supply and demand imbalance piece of the outlook.
Typically, in a mature market, sales would more closely track incomes and inflation or the cost of production. However in the current marijuana market this is not happening due to the ongoing price declines, a result of increased competition. These dynamics are bad news for firms trying to operate a profitable business.
Now, an initial supply response occurred this past year. Total harvest in 2022 declined 13 percent compared to 2021, with an even larger 19 percent decline during peak harvest season. That said, the market still is not in balance. Some of our advisors indicated another similar decline this year, bringing harvest closer to 2019 or 2020 amounts would likely bring the market into better balance. While early in the year always sees the lowest harvest volumes, harvest continues to be down year-over-year in recent months.
We are also starting to see businesses adjust to the market in terms of employment. At OLCC licensed firms that our friends at the Employment Department are able to track in the unemployment insurance system, jobs are down 8 percent year-over-year. Much of this decline is in the agriculture and manufacturing sectors which is where a reduction in harvest and production is likely to be felt the most. Employment in trade, transportation, and warehousing (likely mostly retail) is down the last couple of quarters but unchanged on a year-over-year basis.
The other source of market balance could come from increased consumer demand. That said the low-hanging fruit for demand growth is behind us. Marijuana usage rates are steady in recent years, after increase considerably in the past decade. Many former black market consumers have converted to the legal market, and those that remain may be harder to switch. And underlying population growth has slowed during the pandemic, with only a modest rebound expected in the outlook. [In the just-released 2021 SAMHSA data, marijuana usage rates in Colorado, Oregon, and Washington all held steady at around 19-20 percent.]
Overall, expectations are the market will stabilize in the not-too-distant future. Sales and tax collections will remain relatively steady this year and next. Sales did perk up in March, an indication a sales bottom may have been reached. Overall revenue and resources will be unchanged from the current 2021-23 to 2023-25 biennium. As supply and demand are expected to get into better balance, some pricing power and profitability will return to the market. Overall sales and taxes will increase with a growing population and economy in the decade ahead. Usage rates and consumption as share of income are expected to hold steady in the longer run. Ultimately our office expects marijuana to be a normal good, where sales and consumption generally rise with income gains, even if that has not been the case in the past year due to the price declines. Revenue upside and downside risks primarily lie with what happens to prices, and not so much from an economic/income/population perspective today.
This post is largely based on our office’s most recent forecast, along with updated data where available.
Early in the pandemic when working-from-home increased substantially the concept of Zoom Towns — coined by Bloomberg’s Conor Sen — emerged. These were generally smaller, scenic areas that were desirable places to live and where workers could telecommute from. Zoom Towns garnered a lot of attention, rightfully so, and home sales boomed more than in most other markets. However a key question at the time was whether WFH migration would continue to double down on existing growth patterns, or if new patterns would begin to emerge.
Untangling all of the causes and effects is challenging, but as discussed the other day, we know that many medium sized metros and rural counties have seen faster population gains during the pandemic than they did pre-pandemic. And if we focus specifically on some Zoom Towns in western states, an interesting pattern emerges. Zoom Towns experienced population growth and increased demand. Net domestic migration rates to places like Bend, Bozeman, Missoula, and Spokane were all positive in recent years and people moved in. However, it’s actually the Zoom Towns’ neighboring counties that grew the fastest, as seen in the latest edition of the Graph Map of the Week.
So what’s going on here? I see three issues to highlight.
First would be these surrounding areas are small. These counties have thousands, maybe tens of thousands of residents. So from a net migration perspective we are talking about hundreds of new residents, which in percentage terms is very large, if not necessarily from a raw numbers perspective. Plus we need to recognize potential sample size and margin of error issues with smaller areas.
Second, the uptick in growth could be the result of farther-flung households moving into rural communities. I think this is what most people were wondering earlier in the pandemic, in terms of WFH opportunities boosting growth. Households now untethered from big cities would repopulate the countryside. Something like that. And right now we do not have the data to confirm nor deny that is the case. But, this brings me to the third item which is what I suspect is primarily happening.
Third, the stronger growth in the counties surrounding the Zoom Towns is likely due to regional dynamics and cascading migration. The increased demand of and migration to the Zoom Towns literally or effectively pushed some residents into nearby communities, probability in large part due to housing affordability and availability.
Now, this isn’t necessarily all bad news, or entirely about housing. Some households do prefer to live in smaller towns. As the bigger hub in the region grows, some want to move away, but also remain close enough to maintain community/economic/social ties they have built up over the years. And for the surrounding communities, the stronger growth boosts their local economies as well. Plus we know most rural counties have faced natural population decline in recent decades where deaths outnumber births. Stronger in-migration helps balance local demographics, which will also support a stronger economy and better funded public services in the years ahead.
However, to the extent this cascading migration is about housing affordability it results in economic displacement. Households try to find the right balance in terms of location, opportunity, and affordability. Economists call this constrained optimization. But we know some of our neighbors, particularly our lower- and middle-income friends, family and neighbors, are financially forced to move at times. As such this data highlights the importance of regional housing markets and continued need to increase Oregon’s housing supply. The new Oregon Housing Needs Analysis (OHNA) is aimed to highlight and address these issues, and our office is part of that process in estimating the need and production targets.
For now we do not have any true details on the 2022 data, but given past years and the underlying dynamics, I strongly suspect the cascading migration is the primary driver of these patterns, as opposed to the bigger metro residents moving directly to smaller, rural towns. This will be something I am very interested in seeing as we get more data. Unfortunately, something like the IRS county-to-county migration data comes out with an even longer lag than the Census data. But will be well worth exploring when available.
A few final notes on the Zoom Towns.
Crook County (Prineville) grew the fastest in the state in the new 2022 Census estimates. At least in years past we know a lot of the migration into Crook came from Deschutes. I suspect that remains the case, but we need more data to confirm. That said Crook County’s economy has been booming as well with strong job growth in recent years. From February 2020 to February 2023, Crook County’s employment is up 12%, ranking second fastest among all counties in the state. Statewide employment is up 1% over the same time period. And while being close to Bend, Redmond, the airport, and the like is beneficial, it is clearly not just those things driving Crook’s gains.
Another item that stands out here is how counties are an imperfect unit of measurement. Obviously La Pine and Redmond, which have been growing fast, are located in Deschutes, and somewhere like Belgrade, MT has been booming just outside Bozeman but is also in Gallatin County as well. So focusing just at the county level, which is the data we have available, can mask some of these cascading migration effects which are likely even more pronounced than the map indicates.
And lastly I cannot pretend to know the local dynamics of Jackson, WY. I do not know why it has a negative domestic migration rate during the pandemic. But if you have ever been there the spillover effects and cascading migration is evident. As you head over the pass into Idaho (Teton County) the growth and changes are very apparent. Towns in there like Victor, ID are growing, changing, and housing affordability is dropping as workers and households move 25 miles away from the main city.
No, I’m not trying to be clickbaity, or bump our SEO here, but I have now had a few days to begin to digest the 2022 county population estimates that Census released last week. Here are a few things that stand out to me so far. Keep in mind in Oregon we look at 2 sets of population estimates, one from Portland State’s Population Research Center and the other from the Census Bureau. And also there is some truth to the old adage that demography is destiny. At a base level, how many people live in your community impacts the economy, the labor market, and public revenues. These numbers matter for the outlook.
First, at a high level the 2022 population estimates look a lot like 2021 nationwide. There is still a clear urban-suburban-rural dynamic. I hope you saw this cool domestic migration map Census made for the release last week. You can see the urban outflows not just from Denver, Portland, and Seattle, but also from Atlanta, Dallas, Houston, Memphis, Miami, Nashville, Raleigh and so on.
Second, that said there are some key differences between the 2021 and 2022 patterns. In general, it remains true that most urban cores nationwide continued to lose population last year. However for at least half of the large metros for which county level data is decently useful, 2022 was less bad. The outflows were smaller. That was not the case in Portland (Multnomah County) which saw about the same negative domestic migration rate last year as the year before.
Third, one other difference is in many of those Sun Belt metros their suburbs did grow even as their urban cores declined. That was not the case in the Census estimates for the West Coast. And in the Portland region, Census estimates both Clackamas and Washington counties lost population. That’s different both relative to PSU’s estimates of small gains in the main suburbs, and Census’ 2021 estimates of growth there as well. Clark County is the local exception, with it’s continued growth and net in-migration. Now, I will say Clark’s net domestic migration in 2022 was less than half that of the 2021 estimates (2,300 vs 5,700). As such the impact on the Oregon and Washington statewide numbers was a bit more muted than I had initially expected. It could be growth did slow, or there was also an uptick in out-migration from Clark that lowered the net numbers. We need to see more data to know.
Fourth, one question I had ahead of the data was where and how the Census estimates would differ from the PSU ones. At a high level, Census was more pessimistic pretty much statewide with the things like the population losses in the Portland suburbs and a big slowing in Deschutes. Now, focusing purely on the year-to-year estimates can be a bit problematic due to the noise. But if we step back and look at the 2 years of the pandemic data overall you can see that PSU and Census are broadly in agreement in terms of changes from 2020 to 2022. Yes, Census sees slower growth for most counties, but Census also sees faster growth for 9 counties in the state, as seen in the blue bubbles in below.
Fifth, while it is important to focus on the big picture of the slower growth in Oregon during the pandemic, it’s not uniformly true that everywhere slowed down during the pandemic. 1 in 3 U.S. counties saw faster annual population growth in the past two years than they had averaged in the past couple of decades. These counties — including 11 in Oregon — are disproportionately those located in medium-sized metro areas (250,000 to 1 million population) and rural/nonmetro counties, especially those adjacent to metro areas. This is something I had noticed earlier this year when digging into both the Klamath and Malheur data, and it is even broader than that and across most of Eastern Oregon. This relatively faster growth can have big impacts at the local level in our communities across the state, even if the statewide numbers are slower.
Quick note: Our friends out in Connecticut recently switched from having 8 counties in their state to having 9 “planning regions”. This makes any county or historical work nationwide problematic. For now there is no data for the Constitution State in this map as a result.
As we continue to get more data, our office will work on updating our forecasts accordingly. We still expect population growth to rebound in 2023 and 2024. The real test will be the 2023 population estimates released by PSU this November, and by Census a year from now. The 2022 estimates, which are really July 1, 2021 to July 1, 2022, were still impacted by the pandemic. But another year of stagnant population could really start to make us rethink the longer-term outlook. Now, demographic forecast adjustments tend to be more gradual over time. Something like incorporating a lower birth rate takes time as it continues to drop a little bit each year. But one possibility our office could do would be to develop a full-fledged alternative scenario to gauge how a stagnant population would impact the economy and revenues in the decade ahead. We shall see what the future data holds and our advisors say. Also, stay tuned, I still have one more piece on Zoom Towns and pandemic migration coming.
On Thursday (March 30th) Census will release nationwide population estimates at the county level. We already have Portland State’s county estimates for Oregon, and state estimates from Census. We know the Census data shows net out-migration from Oregon, and we know the PSU estimates show Multnomah is the biggest source of weakness.
I am out of town on Spring Break, so will be unable to respond as the data is released. I will dig into them when I return next week. But, there are a few specific things I am watching for in the data and I thought I would share them in advance of the actual data. Keep in mind one of the biggest benefits of the Census estimates are they are nationwide, so one can compare relative patterns across the nation to get a gauge on what is happening.
First, counties are an imperfect geographic unit to analyze. Counties actually work pretty well in Oregon, especially in the Portland region where the primary (largest) city is mostly contained in a single county, and the other counties are the suburbs. That is far from the case elsewhere in the nation. Large metro areas like Phoenix and San Diego are entirely contained within a single county, and the Los Angeles MSA is just two counties. These differences complicate any analysis trying to get at what happened during the pandemic, and possibly post-pandemic as well, time will tell.
Second, what happened during the pandemic, at least in the 2021 data, is a clear urban-suburban-rural pattern. Across nearly all large metro areas where county level data provides useful information, the urban core county lost population. This wasn’t just in Portland and Seattle, but also in some Sun Belt metros like Dallas, Nashville, and Orlando. Does this relative pattern continue in the 2022 population estimates? Do some of these cities see a rebound? If so, which ones saw relative improvements and which ones saw further losses? That’s a key item to watch.
Third, beyond the urban core story, are there broader patterns in the suburbs or rural areas that are of note? Are some entire metro areas seeing slow/no growth or are the suburbs holding up well even if urban cores are weak? Do we see growth holding steady or even accelerating into rural counties, or some sort of slowing? What about Deschutes County, which has still been growing faster than the state, but growing relatively slower than its historical rate? Does Census also estimate that places like Klamath and Malheur are seeing comparatively stronger gains? Lots of items to check in on with the new data.
Fourth, how does Clark County specifically impact the Oregon and Washington numbers? We know Census estimates the State of Washington also experience net out-migration and we have a good idea that Clark County saw net in-migration, so how does that impact the Portland regional numbers, and then what about Washington ex Clark in terms of migration, especially from the Puget Sound region?
Fifth, there are no silver linings when more people are voting with their feet saying they don’t want to live in your community. Even so, in terms of the economic impacts moving forward, the composition of the changes does matter. It’s not good to lose young families and retirees, but losing working-age households as well is even more problematic in terms of the economy. Now, we will not get the socio-economic characteristics of migrants until this fall when Census releases the 2022 American Community Survey data. But if we take a look at the available data, we see that even in 2021 there was net in-migration to Multnomah County among 20- and 30-somethings. That means young adults were still moving in during the pandemic, and will be starting to set down roots and the like. What does this look like in 2022 and beyond? That will be a key item to watch when the data becomes available.
Looking forward our office’s statewide forecast does call for a modest rebound in population growth and migration in 2023 and 2024. The regional economy is strong and job opportunities are plentiful. It is a very modest rebound, with statwide population gains being less than 1 percent annually for the decade ahead. We know the upcoming 2022 data will be down, we just don’t know exactly what Census estimates the geographic patterns or socio-economic characteristics are yet, that’s what new information we will receive this week and this fall respectively. As we learn more, we can continue to update our thinking on the impacts, the implications, and the forecast. Happy Spring Break everyone.
Mark McMullen
Josh Lehner
Oregon Office of Economic Analysis 